Automation

To the Editors:

Daniel Bell’s review in your August 26 issue admirably disposes of the notion that the rates of change in technology and automation debar or make excessively difficult the attainment and maintenance of maximum employment and minimum unemployment in the United States. I do think that the rate of economic growth required may be some-what higher than Dr. Bell would seem to imply, because the rather long-term average rates of productivity advance which he cites seem to me less relevant than the considerably higher rates of productivity advance in key industries and elsewhere during the very most recent years, which I think are likely to persist or accelerate further.

I agree also with Dr. Bell that an adequate rate of economic growth, which means growth in total demand, would result, with relatively unimportant exceptions, in corresponding modifications in the structure of the labor force, which means that the problem of “structural unemployment” has been grossly exággerated. But Dr. Bell perhaps does not stress sufficiently that vast changes are needed in what I call the structure of demand, or the internal composition of aggregate demand, in that the extent to which a given expansion in demand and production will increase employment depends greatly upon the rates of productivity advance in the particular areas of production against which the demand is exerted. These rates of advance in productivity vary greatly in various sectors of the economy, and also there are wide variations as to the amount by which demand for various products may be expanded in terms of the needs and desires of the people, even assuming that they have plenty of money to spend. This observation supports the very important conclusion, reached by Dr. Bell on value judgments which I devoutly share, that the changes in the structure of demand required to restore and maintain maximum employment require very much more emphasis upon expanded public spending, and a halt to the veritable orgy of tax reduction in which we have thus far engaged. We cannot, with a continuation of this horrible over-reliance on tax reduction as the main weapon of policy, clear our slums, rebuild our cities, meet our grossly under-serviced educational and health needs, replenish and expand our natural resource base, bring our levels of Social Security and welfare payments up to the standards justified by our productive capabilities, nor mount anything approximating a full-scale war against poverty.

I would modify somewhat Dr. Bell’s statement that the net growth rate of 5.5 percent or 6 percent required to achieve the Eisenhower Commission’s goals is “something that is far beyond the capacity of the economy as it is now organized.” Fundamental changes in the economy as it is now organized, as this phrase would commonly be interpreted, are not essential nor even called for. We do need, above all, very substantial reformulation of national economic policies, both from the economic viewpoint (as narrowly defined) and the least equally important viewpoint of ultimate value judgments, along the lines which I have just set forth and with which Dr. Bell seems clearly to agree.

Leon H. Keyserling

Washington, D.C.

Daniel Bell replies:

I think Mr. Michael is right about popular expositions, which is why his pamphlet on Cybernation has always been so inadequate a ground for argument. I am sorry that Mr. Michael did not choose to confront the questions I raised but has simply waffled so that I, for one at least, do not know what he is talking about. Let me restate the issue for the reader.

Mr. Michael has been one of a number of writers who have put forth an apocalyptic vision of the future as a result of computer technology. He predicts a radical break in the economic effects of automation from previous technological change. He argues that technological change is accelerating, and in consequences vast disruptions of the society are in prospect. My criticism of Mr. Michael (and the Ad Hoc Committee on the Triple Revolution) was that he had confused changes in particular industries with aggregate or total effects. If one looked at the only measure available to both of us, inadequate as I tried to show it is, the productivity index, there was no evidence for his case. Automation can be shown to have drastic economic effects only if productivity is rising rapidly (i.e., greater output with fewer workers); it has not. Now he cites Mr. Fabricant about the limitations of the instrument, which I had emphasized. But if he had followed Mr. Fabricant further (in Technology and Social Change, edited by Eli Ginzberg) he would have read: “In order to measure the rate of technological change for the economy as a whole, probably the best measure, although not a good measure—is the productivity index…. If we use it as a measure of technological change we will find no acceleration, no revolutionary change. The rate of increase of productivity in the United States today is of the same order of magnitude as the long term rate of growth; it is approximately 2.5 percent” (p. 115).

Nor is there, unfortunately, an outlook in the near future for such an acceleration. I cited a Labor Department study. I don’t know who Mr. Michael consulted. (The “All” is quite slippery, indeed.) He need only look at the study itself to see its terms of reference. It is to describe “major new machines, processes and products which are believed likely to have an important effect on industry’s unit labor requirements over the next ten years.” What other kind of data could this yield if not the probable effects of automation? I repeat, of 36 industries studied, there is no reference to computers as being likely of use in 16 industries, and for six there is reference only to office applications.

When Mr. Michael, in reply, says that “no available model” can predict the rate at which any industry will cybernate, this is a sorry basis for his prediction of coming cybernation. I tried to use availability of technological processes; and to these can be added costs of investment and replacement as a basis of assessment. But Mr. Michael talks of the “interplay of economic, humanitarian and public relations and other factors” as making cybernation “more or less possible and pervasive” (what waffling!) and says that mine are the “incautious speculations.” (Quel chutzpah.)

In any event, let me anticipate a forthcoming report of the President’s Commission on Automation. The evidence that has been assembled from additional studies on future technology and from a sampling of expert opinion, indicates that no technology not already introduced is likely to have a substantial impact on the economy before 1975, and no presently unknown technology (because of such problem as costs) is likely to be important before 1985.

What Mr. Michael has failed to see is the simple distinction utilized by Herbert A. Simon in his recent book The Shape of Automation (which appeared after my essay had been published and which I recommend here as the best short study on the subject). Mr. Simon calls himself a technological radical but an economic conservative. By this he means that (against some skeptics) he does believe that computers will be capable of revolutionary achievements in areas of decision-making and problem-solving. (To be popular, he believes they can become “thinking machines.”) But he is conservative about its disruptive economic effects. The reasons go back to Adam Smith and to John Maynard Keynes, in the doctrine of comparative advantage in the one, and level of demand in the other. Technically, the decision to use any new technology depends upon comparative price—of labor and capital. The market price, which determines relative use, will be proportional to the marginal productivity of each factor. In other words the structure of employment will shift in proportion to the man-machine productivity ratios. As Simon puts it, “if computers are a thousand times faster than bookkeepers in doing arithmetic, but only a hundred times faster than stenographers in taking dictation, we shall expect the number of book keepers per thousand employees to decrease but number of stenographers to rise.”

The crucial point is that if as a result of automation we see fewer employees in a factory or office, it means actually fewer per unit of output and fewer per unit of capital equipment, but not necessarily fewer in total. The question of the total is a function of the level of demand.

That simple point makes Mr. Michael’s quotation from Mr. Seidman so puzzling. Mr. Seidman is speculating in 1963 about disruptions. But the Government survey that I cited, released in December 1964, is a simple check on such speculations. It showed that of 1,325,000 employees in ten Federal agencies who were affected by the introduction of data processing machines, only 1600 employees, over a three year period, were “displaced” and of these only 2 per cent lost jobs. The reason, simply, was that the demand for government personnel was steady. In the insurance industry all large and medium-size carriers have installed computers. B. L. S. studies show a rise in productivity and in employment in that industry.

The problem of demand is the nub of the issue. Those who talk of automation and cybernation have raised a bogey which simply beclouds the real economic and political problem: how to maintain sufficient demand so as to assure full employment and how to satisfy the evident large needs of major groups in the society. The bald notion that we now live in a “push-button” economy which can turn out abundance at will is fallacious. (If we were to utilize all plant to full capacity today it is unlikely that we could increase G.N.P. more than 7-8 per cent, and even then since the additional equipment, and labor, is probably marginal, it would be at higher costs per unit item. Nor is it a matter either of converting the defense budget. Which accounts for less than 10 per cent of G.N.P.) As I pointed out, if one “costs out” the modest goals proposed by the Eisenhower Commission, which the National Planning Association has done, we do not have the resources to achieve all these goals at the same time (the net deficit, according to the N.P.A. would be $150 billion a year). Certainly, with so many needs to be filled, neither technology nor cybernation is going to make human labor redundant.

The question is one of social priorities. And that is why I find Mr. Michael’s approach so suspect, for what he does is substitute blowsy rhetoric for thought. That the political process is a difficult one indeed need not be underscored. But such apocalyptic statements as “we have to transform our values” makes no sense in relation to full employment. (What it does, perhaps, is to reinforce the emotions of young radicals that nothing is possible, and thus their rhetoric.) If Mr. Michael looked about the world he would see that for the last ten years, without war, western Germany, France, England, Switzerland, Belgium, Holland, the Scandinavian countries, Japan have had full employment, with labor shortages to boot, and that the United States, in the years after World War II, had full employment even while dismantling a large-scale military apparatus. We lost full employment in the U.S. when the Eisenhower administration chose price stability over economic growth and the slowdown of growth combined with the tide of young people coming into the labor force created employment deficits that are only now slowly being reduced.

Full employment is no problem, without war, in any modern economy where the government is willing to use the fiscal powers available to it. In fact, “full employment” does represent now the basic value committment of the government. What is more difficult, however, is the achievement of a different social pattern which suggests for example, that increased revenues be used to transform the cities rather than be returned as tax rebates which increase the purchases of automobiles. But Mr. Michael shows little faith in the possibility of change. He wrings his hands and waits only for “great transformations.” I am not the one who is lamenting. It is he who laments. It is Mr. Michael and his friends who wail, for they stand on Mount Pisgah and do not know how to come down.